Predrilling formation pressure prediction method, system, equipment and medium
By building a formation pressure prediction system based on LSTM and Eaton models, high-precision pressure prediction of undrilled sections is performed using drilled section data, which solves the problem of inaccurate prediction in traditional methods and improves drilling safety.
Patent Information
- Application Number
- CN202510974154.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-15
- Publication Date
- 2025-09-09
AI Technical Summary
Traditional pre-drilling formation pressure prediction methods rely on qualitative analysis and empirical formulas, which makes it difficult to achieve accurate predictions in complex drilling environments and cannot meet high-precision early warning requirements.
By acquiring logging, mud recording and seismic data of the drilled section, the LSTM model and Eaton model are used to construct a formation pressure prediction model. The well depth data of the undrilled section is collected in real time for prediction, and high-precision prediction is performed by combining the layer velocity and formation pressure model.
It achieves high-precision real-time prediction of formation pressure, is suitable for complex working conditions such as high-temperature and high-pressure wells and deep-water wells, and improves the well control safety of drilling operations.
Smart Images
Figure CN120608676A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of drilling safety assurance, and in particular to a method, system, equipment and medium for predicting formation pressure before drilling. Background Art
[0002] Predicting formation pressure helps improve the safety factor of drilling, thereby facilitating a safe and economical drilling process. Accurate formation pressure prediction is particularly important in new exploration areas or under unknown formation conditions. Traditional pre-drilling formation pressure prediction methods rely primarily on qualitative analysis and empirical formulas, such as determining pressure status by monitoring changes in parameters such as downhole pressure, flow rate, and mud density. However, these methods have significant limitations in data processing and feature extraction, making it difficult to fully exploit the potential sequential relationships among numerous parameters. This results in insufficient prediction accuracy in complex drilling environments, making it difficult to meet the needs of accurate early warning. Therefore, an efficient and accurate pre-drilling formation pressure prediction method is urgently needed to address the above issues. Summary of the Invention
[0003] In view of the above problems, the present invention is proposed to provide a method, system, device and medium for predicting formation pressure before drilling that overcomes the above problems or at least partially solves the above problems.
[0004] To achieve the above-mentioned and other related purposes, the present invention provides a method for predicting formation pressure before drilling, the method comprising:
[0005] Acquire characteristic data of measured points at different formation depths in the drilled section, and determine target parameters closely related to formation pressure from the characteristic data; wherein the characteristic data includes well logging data, mud logging data, and seismic data; the target parameters include well depth and its corresponding formation velocity;
[0006] Extracting well depth data and corresponding interval velocity data corresponding to the target parameter from the feature data, preprocessing the well depth data and the corresponding interval velocity data to obtain preprocessed well depth data and the corresponding interval velocity data, and training a pre-constructed initial LSTM model using the preprocessed well depth data and the corresponding interval velocity data to obtain an interval velocity prediction model;
[0007] A multidimensional data set is constructed using the well depth data and its corresponding interval velocity data, and a formation pressure prediction model is trained by combining the calculation formula of the Eaton model;
[0008] The measured well depth values of the measured points of the undrilled section located in the same address area as the drilled section are collected in real time, and the formation pressure of the undrilled section is predicted in combination with the layer velocity prediction model and the formation pressure prediction model to obtain the predicted value of the formation pressure of the undrilled section.
[0009] Optionally, determining a target parameter closely related to formation pressure from the characteristic data includes:
[0010] Performing Spearman correlation analysis on the well logging data, the mud logging data, and the seismic data to calculate the rank correlation coefficient between each parameter in the well logging data, the mud logging data, and the seismic data and the formation pressure;
[0011] The parameter corresponding to the rank correlation coefficient greater than the preset rank value is determined as the target parameter.
[0012] Optionally, preprocessing the well depth data and the corresponding interval velocity data to obtain the preprocessed well depth data and the corresponding interval velocity data includes:
[0013] performing data cleaning on the well depth data and the corresponding interval velocity data to obtain cleaned well depth data and the corresponding interval velocity data;
[0014] Normalization processing is performed on the cleaned well depth data and the corresponding interval velocity data to obtain pre-processed well depth data and the corresponding interval velocity data.
[0015] Optionally, the pre-constructed initial LSTM model is trained using the pre-processed well depth data and its corresponding interval velocity data to obtain an interval velocity prediction model, including:
[0016] The preprocessed well depth data and its corresponding interval velocity data are divided into a training set and a test set;
[0017] The well depth data preprocessed in the training set is used as the input of the initial LSTM model, and the interval velocity preprocessed in the training set is used as the output of the initial LSTM model. The initial LSTM model is iteratively trained until the training is stopped when the preset number of iterations is reached, and the corresponding interval velocity prediction model is obtained.
[0018] Optionally, after the step of iteratively training the initial LSTM model until a preset number of iterations is reached and the training is stopped to obtain the corresponding layer velocity prediction model, the method further includes:
[0019] Inputting the true value of the well depth in the test set into the interval velocity prediction model, and outputting the interval velocity test value corresponding to the true value of the well depth;
[0020] Obtaining a true value of interval velocity corresponding to the true value of the well depth from the test set, and calculating a mean absolute error and a root mean square error of the interval velocity prediction model using the true value of the interval velocity and the interval velocity test value;
[0021] The model performance of the interval velocity prediction model is evaluated using the mean absolute error and the root mean square error.
[0022] Optionally, the well depth data and its corresponding interval velocity data are used to construct a multidimensional dataset, and the formation pressure prediction model is trained in combination with the calculation formula of the Eaton model, including:
[0023] Calculating various parameters for training the formation pressure prediction model based on the well depth data and its corresponding interval velocity data, and constructing a multidimensional data set using the various parameters;
[0024] According to the multidimensional data set, the calculation formula of the Eaton model is used as a basic model for training to obtain a formation pressure prediction model.
[0025] Optionally, the real-time acquisition of measured well depth values of measured points in an undrilled section located in the same address area as the drilled section, and prediction of the formation pressure of the undrilled section in combination with the interval velocity prediction model and the formation pressure prediction model to obtain a predicted formation pressure value of the undrilled section, includes:
[0026] Real-time acquisition of measured well depth values of measured points in an undrilled section located in the same address area as the drilled section, inputting the measured well depth values into the interval velocity prediction model, and outputting interval velocity prediction values corresponding to the measured well depth values;
[0027] According to the measured well depth value and the corresponding predicted layer velocity value, the formation pressure of the undrilled well section is predicted using the formation pressure prediction model to obtain a predicted formation pressure value.
[0028] In a second aspect, the present invention further provides a pre-drilling formation pressure prediction system, the system comprising:
[0029] a determination module, configured to obtain characteristic data at a measured point of formation pressure in a drilled well, and determine a target parameter closely related to the formation pressure from the characteristic data; wherein the characteristic data includes well logging data, mud logging data, and seismic data; and the target parameter includes well depth and its corresponding formation velocity;
[0030] a preprocessing module, configured to extract the well depth data and the corresponding interval velocity data corresponding to the target parameter from the feature data, preprocess the well depth data and the corresponding interval velocity data to obtain the preprocessed well depth data and the corresponding interval velocity data, and train a pre-built initial LSTM model using the preprocessed well depth data and the corresponding interval velocity data to obtain an interval velocity prediction model;
[0031] A training module is used to construct a multidimensional data set using the well depth data and its corresponding interval velocity data, and to train a formation pressure prediction model in combination with the calculation formula of the Eaton model;
[0032] The prediction module is used to collect the measured well depth values of the measured points of the undrilled section located in the same address area as the drilled section in real time, and predict the formation pressure of the undrilled section in combination with the layer velocity prediction model and the formation pressure prediction model to obtain the predicted value of the formation pressure of the undrilled section.
[0033] In a third aspect, the present invention provides an electronic device comprising: a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program stored in the memory, so that the electronic device performs the steps of the pre-drilling formation pressure prediction method as described above.
[0034] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by an electronic device, implements the steps of the pre-drilling formation pressure prediction method as described above.
[0035] The above one or more technical solutions provided by the present invention may have the following advantages or at least achieve the following technical effects:
[0036] The present invention provides a pre-drilling formation pressure prediction method, system, equipment, and medium. The method utilizes well logging data, mud logging data, and seismic data from measured points at different formation depths in a drilled section for correlation analysis to determine the well depth and layer velocity, which are closely related to formation pressure. The well depth and its corresponding layer velocity are then used to train a pre-built initial LSTM model to obtain a layer velocity prediction model. Furthermore, a multidimensional dataset is constructed using the well depth and layer velocity, and combined with the calculation formula of the Eaton model, a formation pressure prediction model is trained. The well depth measured values of measured points in an undrilled section located in the same address area as the drilled section are collected in real time, and the formation pressure of the undrilled section is predicted using the layer velocity prediction model and the formation pressure prediction model to obtain a predicted value for the formation pressure of the undrilled section. This method achieves high-precision, real-time prediction of formation pressure. This method is applicable to complex working conditions such as high-temperature, high-pressure wells and deepwater wells, providing more accurate pressure assessments for drilling operations and improving well control safety. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] Figure 1 Shown is a flow chart of a method for predicting formation pressure before drilling in one embodiment of the present invention;
[0038] Figure 2 Shown is a schematic diagram of a process for predicting formation pressure in one embodiment of the present invention;
[0039] Figure 3Shown is a schematic diagram of the structure of an initial LSTM model pre-built in one embodiment of the present invention;
[0040] Figure 4 Shown is a schematic diagram of functional modules of a pre-drilling formation pressure prediction system according to one embodiment of the present invention;
[0041] Figure 5 Shown is a schematic diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0042] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. The details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless they conflict.
[0043] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.
[0044] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it will be apparent to those skilled in the art that the embodiments of the present invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the embodiments of the present invention.
[0045] In the description and claims of the embodiments of the present disclosure, as well as in the accompanying drawings, the terms "first," "second," and the like are used to distinguish similar items and are not necessarily used to describe a particular order or precedence. It should be understood that such terms are interchangeable where appropriate for the purposes of describing the embodiments of the present disclosure herein. Furthermore, the terms "including," "having," and any variations thereof are intended to cover non-exclusive inclusions.
[0046] Unless otherwise stated, the term "plurality" means two or more.
[0047] In the embodiments of the present disclosure, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B means: A or B.
[0048] The term "and / or" describes an association between objects, indicating that three relationships can exist. For example, "A and / or B" means: A or B, or A and B.
[0049] The technical solutions in the embodiments of the present invention will be described in detail below with reference to the accompanying drawings in the embodiments of the present invention.
[0050] See also Figure 1-2 An embodiment of the present invention provides a method for predicting formation pressure before drilling, which may include the following steps S10 to S40:
[0051] Step S10, obtaining characteristic data of measured points at different formation depths in the drilled section, and determining target parameters closely related to formation pressure from the characteristic data; wherein the characteristic data includes well logging data, mud logging data and seismic data; the target parameters include well depth and its corresponding layer velocity.
[0052] Characteristic data represents multiple parameters associated with formation pressure and can include well logging data, mud logging data, and seismic data. For example, parameters associated with formation pressure can include overburden pressure, mud density, pump pressure, flow rate, seismic reflection time difference and velocity, seismic impedance, and reflection amplitude.
[0053] Well logging data may include rock formation resistivity, porosity, oil and gas content, well depth, etc.
[0054] Well logging data may include overburden pressure, mud density, pump pressure, flow rate, and formation pressure.
[0055] Seismic data may include seismic reflection moveout and velocity, seismic impedance, reflection amplitude, and (stratum) layer velocity.
[0056] Target parameters are used to represent parameters closely related to formation pressure; they may include: well depth and its corresponding layer velocity.
[0057] In a specific implementation, characteristic data of measured points at different formation depths in the drilled section can be obtained first; the characteristic data can include well logging data, mud logging data and seismic data; then, correlation analysis can be performed between each parameter in the characteristic data and the formation pressure (such as principal component analysis, Spearman correlation analysis, etc.) to screen out one or more target parameters that are closely related to the formation pressure; the target parameters can include the well depth and its corresponding layer velocity.
[0058] As an example, principal component analysis (PCA) is used to calculate the variance contribution of each parameter in the feature data relative to formation pressure. The variance contribution rates of each parameter are then sorted from largest to smallest to form a parameter sequence. The first N parameters in this sequence are then identified as target factors, where N is a positive integer; N can be set based on actual conditions. The variance contribution rate can be used to characterize the correlation between each parameter and formation pressure. A higher variance contribution rate for a parameter indicates a higher correlation with formation pressure. Parameters with variance contribution rates exceeding a preset variance contribution threshold can also be identified as target parameters. This allows PCA to filter multiple target factors from multiple parameters in the feature data.
[0059] Step S20: extracting the well depth data and the corresponding interval velocity data corresponding to the target parameter from the feature data, preprocessing the well depth data and the corresponding interval velocity data to obtain the preprocessed well depth data and the corresponding interval velocity data, and using the preprocessed well depth data and the corresponding interval velocity data to train a pre-built initial LSTM model to obtain an interval velocity prediction model.
[0060] The pre-processed well depth data may be the well depth obtained after a pre-processing operation is performed on the well depth data.
[0061] The preprocessed layer velocity data may be the layer velocity obtained after performing a preprocessing operation on the layer velocity data.
[0062] The initial LSTM model can be an LSTM model with a pre-built neural network structure. By utilizing the LSTM model's ability to process time-series feature data, the present invention can capture hidden pressure variation patterns in feature data (i.e., well logging data, mud recording data, and seismic data). Furthermore, based on the well depth data of the drilled section and its corresponding interval velocity data, the present invention can predict the interval velocity corresponding to the well depth when the well depth of the undrilled section located in the same address area as the drilled section is known.
[0063] See also Figure 3 , Figure 3The diagram shows the structure of the initial LSTM model. This initial LSTM model consists of an input layer, a hidden layer, and an output layer. The input layer has two neurons, representing the well depth (or formation depth) and its corresponding layer velocity (the current layer velocity corresponding to the formation depth). Two hidden layers are configured, with five and four neurons, respectively. At each time step, the LSTM unit updates its internal state based on the previous state, the current input, and gate control, and inputs the current hidden state. Therefore, the LSTM model can both retain long-term historical information and flexibly respond to the latest input, adapting to the modeling needs of complex time series data and minimizing overall error. The output layer's neurons are configured to represent the layer velocity predictions.
[0064] The interval velocity prediction model may be a model obtained by training a pre-built initial LSTM model using pre-processed well depth data and its corresponding interval velocity data.
[0065] In a specific implementation, after determining the target parameters including well depth and layer velocity, the well depth data and its corresponding layer velocity data corresponding to the target parameters can be extracted from the feature data; then, the well depth data and its corresponding layer velocity data are preprocessed, such as data cleaning (outlier deletion, data completion, data replacement, etc.) and normalization, to obtain the preprocessed well depth data and its corresponding layer velocity data; then, the preprocessed well depth data and its corresponding layer velocity data are used to construct a data set, which is divided into a training set and a test set, and the pre-constructed initial LSTM model is trained to obtain a layer velocity prediction model to minimize the overall error in predicting the pre-drilling formation layer velocity.
[0066] Step S30: construct a multidimensional data set using the well depth data and its corresponding interval velocity data, and train a formation pressure prediction model based on the calculation formula of the Eaton model.
[0067] The multidimensional dataset can be a dataset used to train the Eaton model (i.e., the calculation formula of the Eaton model). ob 、p h 、V n , etc.); it may include formation pore pressure p p , normal hydrostatic pressure p h (calculated from the well depth extracted from the characteristic data in step S10 above), overburden pressure (or total overburden pressure) p ob (extracted from the characteristic data of step S10 above), layer velocity V under normal compaction conditions n (extracted from the feature data of step S10 above) etc.
[0068] The Eaton model method uses the acoustic (or other) logging data of the normally compacted formation to establish a normal compaction trend line, and then calculates the formation pressure by the deviation between the actual measurement value and the trend line value. The calculation formula of the Eaton model can be expressed as:
[0069]
[0070] Where p p is the formation (pore) pressure; p h is the normal hydrostatic pressure; p ob is the overburden pressure; V n is the layer velocity under normal compaction conditions; V is the actually measured layer velocity; N is the Eaton index, which is an empirical parameter (dimensionless) whose value usually depends on the specific geological conditions and calibration data. The common value range of N is [1, 3], such as N is 2.7.
[0071] The Eaton model method is used in combination with physical constraints such as rock mechanics and pore fluid dynamics to establish the Eaton model. That is, the Eaton model method is used to use the acoustic (or other) logging data of the normally compacted formation to establish a normal compaction trend line, and then the formation pressure is obtained by the deviation between the layer velocity predicted by the layer velocity prediction model and the trend line value.
[0072] In a specific implementation, the normal hydrostatic pressure p in the calculation formula of the Eaton model can be calculated based on the well depth data and its corresponding layer velocity data for training the formation pressure prediction model (such as h ) and use the various parameters to construct a multidimensional data set; then substitute the various parameters in the multidimensional data set into the calculation formula of the Eaton model, and use the calculation formula as the Eaton basic model for training to obtain a trained formation pressure prediction model.
[0073] Step S40, real-time acquisition of measured well depth values of measured points of an undrilled section located in the same address area as the drilled section, and prediction of the formation pressure of the undrilled section in combination with the layer velocity prediction model and the formation pressure prediction model to obtain a predicted value of the formation pressure of the undrilled section.
[0074] The measured well depth value is used to indicate the actual well depth collected at the measured point in the undrilled section located in the same address area as the drilled section.
[0075] The predicted formation pressure value is used to represent the formation pressure predicted based on the measured well depth in the undrilled section.
[0076] In the specific implementation, after the formation pressure prediction model is trained, the measured well depth values of the measured points of the undrilled section located in the same address area as the drilled section can be collected in real time, and then the measured well depth values are input into the layer velocity prediction model, and the layer velocity prediction value corresponding to the measured well depth value is output; then, based on the measured well depth value and the layer velocity prediction value, the formation pressure of the undrilled section is predicted using the formation pressure prediction model to obtain the predicted formation pressure value of the undrilled section.
[0077] Furthermore, in one embodiment, step S20 may include the following sub-steps S201-202:
[0078] Sub-step S201, performing data cleaning on the well depth data and the corresponding interval velocity data to obtain cleaned well depth data and the corresponding interval velocity data;
[0079] Sub-step S202 : normalizing the cleaned well depth data and the corresponding interval velocity data to obtain pre-processed well depth data and the corresponding interval velocity data.
[0080] In a specific implementation, after extracting the well depth data and its corresponding layer velocity data corresponding to the target parameters, the well depth data and its corresponding layer velocity data can be cleaned to obtain the cleaned well depth data and its corresponding layer velocity data, so as to improve the prediction accuracy of subsequent models (such as layer velocity prediction model, formation pressure prediction model); specifically, the negative values in the well depth data and its corresponding layer velocity data can be replaced with zero; for missing data in any time series data, it can be replaced with the weighted average of the first 10 data values before the missing data; the missing data can also be supplemented by interpolation; and then the cleaned well depth data and its corresponding layer velocity data can be normalized (such as using minimum-maximum normalization, Min-Max Normalization) to obtain pre-processed well depth data and its corresponding layer velocity data, so as to eliminate the dimension and reduce the impact of different orders of magnitude of each parameter on the prediction of subsequent models.
[0081] As an example, the cleaned well depth data and the cleaned interval velocity data corresponding to each well depth data are normalized and scaled to the interval [0, 1] by using the Min-Max Normalization method. This can be achieved by the following formula (2):
[0082]
[0083] Where x ′is the normalized value obtained after normalizing x (such as the normalized value of well depth and the normalized value of interval velocity); x is the original data (the cleaned well depth data and the values in the cleaned interval velocity data corresponding to each well depth data); x max is the maximum value of all cleaned data (such as the maximum value of cleaned well depth data, the maximum value of cleaned interval velocity data); x min It is the minimum value of all cleaned data (such as the minimum value of cleaned well depth data, the minimum value of cleaned interval velocity data).
[0084] Furthermore, in one embodiment, step S40 may include the following sub-steps S401 to S402:
[0085] Sub-step S401, collecting measured well depth values of measured points in an undrilled section located in the same address area as the drilled section in real time, inputting the measured well depth values into the interval velocity prediction model, and outputting interval velocity prediction values corresponding to the measured well depth values;
[0086] Sub-step S402 , predicting the formation pressure of the undrilled section using the formation pressure prediction model according to the measured well depth value and the corresponding predicted layer velocity value, to obtain a predicted formation pressure value.
[0087] In a specific implementation, after the formation pressure prediction model is trained, the measured well depth values of the measured points of the undrilled section located in the same address area as the drilled section can be collected in real time, and the measured well depth values can be input into the layer velocity prediction model to predict the layer velocity corresponding to the measured well depth values, and the layer velocity prediction value can be output; then, based on the measured well depth values and their corresponding layer velocity prediction values, and combined with the Eaton calculation formula, the formation pressure at the measured well depth value of the undrilled section can be predicted using the formation pressure prediction model to obtain the formation pressure prediction value; thereby improving the prediction accuracy and reliability, and achieving high-precision real-time prediction of formation pressure.
[0088] In this embodiment, correlation analysis is performed using well logging data, mud logging data, and seismic data from measured points at different formation depths in the drilled section to determine the well depth and layer velocity, which are closely related to formation pressure. The well depth and its corresponding layer velocity are then used to train a pre-built initial LSTM model to obtain a layer velocity prediction model. Furthermore, a multidimensional dataset is constructed using the well depth and layer velocity, and combined with the calculation formula of the Eaton model, a formation pressure prediction model is trained. The measured well depth values of measured points in the undrilled section located in the same address area as the drilled section are collected in real time, and the formation pressure of the undrilled section is predicted using the layer velocity prediction model and the formation pressure prediction model to obtain the predicted formation pressure value of the undrilled section. This method achieves high-precision real-time prediction of formation pressure. This method is applicable to complex working conditions such as high-temperature and high-pressure wells and deepwater wells, providing more accurate pressure assessment for drilling operations and improving well control safety.
[0089] Based on the above embodiments, a second embodiment of the method for predicting formation pressure before drilling of the present invention is proposed. In this embodiment, step S10 may include the following sub-steps S101 to S102:
[0090] Sub-step S201 : performing Spearman correlation analysis on the well logging data, the mud logging data and the seismic data to calculate the rank correlation coefficient between each parameter in the well logging data, the mud logging data and the seismic data and the formation pressure.
[0091] The rank correlation coefficient ρ can be used to characterize the correlation between various parameters in well logging data, mud logging data, and seismic data and formation pressure. The larger the rank correlation coefficient ρ corresponding to a parameter, the higher the correlation between that parameter and formation pressure.
[0092] In a specific implementation, the Spearman correlation analysis method can be used to perform correlation analysis on the parameters in the well logging data, mud logging data and seismic data and the formation pressure, and then calculate the rank correlation coefficient ρ between the parameters in the well logging data, mud logging data and seismic data and the formation pressure.
[0093] Sub-step S202: determining the parameter corresponding to the rank correlation coefficient greater than the preset rank value as the target parameter.
[0094] The preset rank value may be a preset Spearman rank correlation coefficient threshold. In this embodiment, the preset rank value is preferably set to 0.85.
[0095] As an example, a parameter that reaches a preset rank value (eg, 0.85) may be selected as the target factor.
[0096] It should be noted that the correlation between each target parameter and the formation pressure is higher than the correlation between the remaining parameters and the formation pressure.
[0097] In a specific implementation, after obtaining the rank correlation coefficient corresponding to each parameter, the rank correlation coefficient corresponding to each parameter can be compared with the preset rank value to screen out the rank correlation coefficient greater than the preset rank value; and then the parameter corresponding to the rank correlation coefficient greater than the preset rank value is determined as the target parameter; for example, the target parameters determined by the Spearman correlation analysis method include well depth and layer velocity.
[0098] In this embodiment, by performing Spearman correlation analysis on the well logging data, the mud logging data and the seismic data, the rank correlation coefficient between each parameter in the well logging data, the mud logging data and the seismic data and the formation pressure is calculated; the parameter corresponding to the rank correlation coefficient greater than the preset rank value is determined as the target parameter; considering that drilling is a serialized process that changes with the depth of the formation, the Spearman correlation analysis method can be used to mine the serial relationship between each parameter in the well logging data, the mud logging data and the seismic data and the formation pressure, so as to screen out one or more target parameters closely related to the formation pressure from multiple parameters, thereby providing reliable support for subsequent layer velocity prediction and formation pressure prediction.
[0099] Based on the above embodiments, a third embodiment of the method for predicting formation pressure before drilling of the present invention is proposed. In this embodiment, step 20 may further include the following sub-steps S203 to S204:
[0100] Sub-step S203 : dividing the pre-processed well depth data and its corresponding interval velocity data into a training set and a test set.
[0101] In a specific implementation, a data set can be constructed using the preprocessed well depth data and its corresponding layer velocity data; wherein the data set includes multiple sample data, each sample data includes the normalized value of the well depth and the normalized value of the layer velocity at the same moment (that is, the preprocessed well depth data and its corresponding layer velocity data); and then the sample data in the data set are divided into a training set and a test set, such as 70% of all the sample data constitute the training set, and the remaining 30% constitute the test set.
[0102] Sub-step S204: iteratively train the initial LSTM model using the training set until the training is stopped when a preset number of iterations is reached, thereby obtaining a corresponding layer velocity prediction model.
[0103] The preset number of iterations may be a preset number of training iterations of the initial LSTM model; for example, the number of training iterations may be 50, 100, or 200.
[0104] In a specific implementation, the training set (i.e., the normalized value of each parameter in the training set sample data) can be input into the input layer of the initial LSTM model, and the output layer is set to the predicted value of the (pre-drilling formation) layer velocity (the predicted normalized value of the pre-drilling formation layer velocity); the initial LSTM model is then iteratively trained using the training set until the training is stopped when the preset number of iterations is reached, and the corresponding (pre-drilling formation) layer velocity prediction model is obtained through training.
[0105] Furthermore, in one embodiment, after sub-step S204, the pre-drilling formation pressure prediction method further includes sub-steps S205 to S207:
[0106] Sub-step S205, inputting the true value of the well depth in the test set into the interval velocity prediction model, and outputting the interval velocity test value corresponding to the true value of the well depth;
[0107] Sub-step S206, obtaining the true value of the interval velocity corresponding to the true value of the well depth from the test set, and calculating the mean absolute error and root mean square error of the interval velocity prediction model using the true value of the interval velocity and the test value of the interval velocity;
[0108] Sub-step S207 , evaluating the model performance of the interval velocity prediction model using the mean absolute error and the root mean square error.
[0109] The true value of the well depth is used to represent the normalized value of each well depth in the test set sample data.
[0110] The interval velocity test value is used to indicate the test normalized value of the formation velocity obtained after prediction based on the true value of each well depth.
[0111] The true value of the interval velocity is used to represent the normalized value of the interval velocity corresponding to the true value of each well depth in the test set sample data.
[0112] Mean absolute error (MAE) is used to represent the average value of the absolute error between the true value and the predicted value; it can better reflect the actual situation of the predicted value error.
[0113] Root mean square error (RMSE) is used to measure the deviation between the predicted value and the true value.
[0114] In a specific implementation, after the interval velocity prediction model is trained, the true values of each well depth in the test set can be input into the interval velocity prediction model. The interval velocity prediction model then outputs the interval velocity test values corresponding to each true well depth value. The true interval velocity values corresponding to each true well depth value are then extracted from the test set. The mean absolute error (MAE) and root mean square error (RMSE) of the interval velocity prediction model are then used to calculate the mean absolute error (MAE) and root mean square error (RMSE) of the interval velocity prediction model to verify its performance. The smaller the MAE and RMSE, the higher the prediction accuracy of the interval velocity prediction model.
[0115] In this embodiment, the preprocessed well depth data and its corresponding interval velocity data are divided into a training set and a test set; the initial LSTM model is iteratively trained using the training set until the training is stopped when a preset number of iterations is reached, thereby obtaining a corresponding interval velocity prediction model; thus, the initial LSTM model is trained using the preprocessed well depth data and its corresponding interval velocity data to obtain a (pre-drilling formation) interval velocity prediction model, thereby minimizing the overall error in predicting the pre-drilling formation interval velocity and improving the accuracy of the model prediction.
[0116] Based on the above embodiments, a fourth embodiment of the method for predicting formation pressure before drilling of the present invention is proposed. In this embodiment, step 30 may further include the following sub-steps S301-302:
[0117] Sub-step S301 : calculating various parameters for training the formation pressure prediction model based on the well depth data and its corresponding interval velocity data, and constructing a multidimensional data set using the various parameters.
[0118] In a specific implementation, the parameters in the calculation formula of the Eaton model can be calculated based on the well depth data and its corresponding interval velocity data. For example, the normal hydrostatic pressure p can be calculated based on the well depth data. h , extract the overburden pressure p from the characteristic data of step S10 ob and the layer velocity V under normal compaction conditions n etc.; and then we can use the various parameters in the calculation formula of Eaton model (such as p ob 、p h 、V n etc.) to construct a multidimensional dataset for training the Eaton model.
[0119] Sub-step S302: Based on the multi-dimensional data set, the calculation formula of the Eaton model is used as a basic model for training to obtain a formation pressure prediction model.
[0120] In a specific implementation, after determining the multidimensional data set, the multi-source data set can be substituted into the calculation formula of the Eaton model to obtain the calculated value of the formation pressure, and a normal compaction trend line between the formation pressure calculated according to the Eaton method and the measured formation pressure can be constructed; then, the deviation between the measured formation pressure and this trend curve value can be used to adjust the Eaton model to obtain a formation pressure prediction model; thereby ensuring the prediction accuracy of the formation pressure prediction model.
[0121] In this embodiment, various parameters used to train the formation pressure prediction model are calculated based on the well depth data and its corresponding interval velocity data, and a multidimensional dataset is constructed using these parameters. Based on this multidimensional dataset, the Eaton model's calculation formula is used as a base model for training to obtain a formation pressure prediction model. Thus, based on key parameters such as well depth, interval velocity, and overburden pressure, and combined with the Eaton model's calculation formula, a (pre-drilling) formation pressure prediction model is obtained, thereby improving the accuracy and reliability of pre-drilling formation pressure predictions.
[0122] Based on the same inventive concept, the fifth embodiment of the present invention also provides a pre-drilling formation pressure prediction system corresponding to the pre-drilling formation pressure prediction method of the aforementioned embodiment. Since the principle of the problem solved by the system in the fifth embodiment of the present invention is similar to that of the pre-drilling formation pressure prediction method of the aforementioned embodiment of the present invention, the implementation of the system can refer to the implementation of the method, and the repeated parts will not be repeated. Figure 4 The pre-drilling formation pressure prediction system of the present invention may include:
[0123] The determination module 10 is configured to obtain characteristic data at a measured point of formation pressure in a drilled well and determine a target parameter closely related to the formation pressure from the characteristic data; wherein the characteristic data includes well logging data, mud logging data, and seismic data; and the target parameter includes well depth and its corresponding formation velocity;
[0124] A preprocessing module 20 is configured to extract the well depth data and the corresponding interval velocity data corresponding to the target parameter from the feature data, preprocess the well depth data and the corresponding interval velocity data to obtain the preprocessed well depth data and the corresponding interval velocity data, and train a pre-built initial LSTM model using the preprocessed well depth data and the corresponding interval velocity data to obtain an interval velocity prediction model;
[0125] A training module 30 is used to construct a multidimensional data set using the well depth data and its corresponding interval velocity data, and to train a formation pressure prediction model in combination with the calculation formula of the Eaton model;
[0126] The prediction module 40 is used to collect the measured well depth values of the measured points of the undrilled section located in the same address area as the drilled section in real time, and predict the formation pressure of the undrilled section in combination with the layer velocity prediction model and the formation pressure prediction model to obtain the predicted value of the formation pressure of the undrilled section.
[0127] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which implements the above-mentioned method for predicting formation pressure before drilling when the program is executed by a processor.
[0128] Figure 5 Schematic block diagram of an electronic device provided in an embodiment of the present application. Figure 5 As shown, the electronic device includes: at least one processor 401, a memory 402, at least one network interface 403 and a user interface 405. The various components in the electronic device are coupled together via a bus system 404. It is understood that the bus system 404 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 404 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, Figure 5 Various buses are labeled as bus systems.
[0129] The user interface 405 may include a display, a keyboard, a mouse, a trackball, a click gun, keys, buttons, a touch pad or a touch screen.
[0130] It will be appreciated that the memory 402 may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. The non-volatile memory may be a read-only memory (ROM) or a programmable read-only memory (PROM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM) and synchronous static random access memory (SSRAM). The memories described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memories.
[0131] The memory 402 in this embodiment of the present invention is used to store various types of data to support the operation of the electronic device 400. Examples of this data include any executable program used to operate on the electronic device 400, such as an operating system 4021 and application programs 4022. The operating system 4021 includes various system programs, such as a framework layer, a core library layer, and a driver layer, for implementing various basic services and processing hardware-based tasks. Application programs 4022 may include various application programs, such as a media player and a browser, for implementing various application services. The pre-drilling formation pressure prediction method provided in this embodiment of the present invention may be included in application programs 4022.
[0132] The methods disclosed in the above embodiments of the present invention can be applied to or implemented by processor 401. Processor 401 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits or software instructions in processor 401. Processor 401 may be a general-purpose processor, a digital signal processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc. Processor 401 can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. General-purpose processor 401 may be a microprocessor or any conventional processor. The steps of the pre-drilling formation pressure prediction method provided in conjunction with the embodiments of the present invention can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium located in a memory. The processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.
[0133] In an exemplary embodiment, the electronic device 400 may be configured to execute the aforementioned method by one or more application specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), or complex programmable logic devices (CPLDs).
[0134] In summary, the present invention uses the well logging data, mud recording data, and seismic data of the measured points at different formation depths in the drilled section to perform correlation analysis to determine the well depth and layer velocity that are closely related to the formation pressure; then uses the well depth and its corresponding layer velocity to train the pre-built initial LSTM model to obtain a layer velocity prediction model; then uses the well depth and layer velocity to construct a multidimensional data set, and combines the calculation formula of the Eaton model to train a formation pressure prediction model; real-time acquisition of the measured well depth values of the measured points of the undrilled section located in the same address area as the drilled section, and combines the layer velocity prediction model and the formation pressure prediction model to predict the formation pressure of the undrilled section, thereby obtaining the formation pressure prediction value of the undrilled section; thereby achieving high-precision real-time prediction of formation pressure. This method is suitable for complex working conditions such as high-temperature and high-pressure wells and deepwater wells, providing more accurate pressure assessment for drilling operations and improving well control safety.
[0135] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.
Claims
1. A method for predicting formation pressure before drilling, characterized in that: The method comprises: Acquire characteristic data of measured points at different formation depths in the drilled section, and determine target parameters closely related to formation pressure from the characteristic data; wherein the characteristic data includes well logging data, mud logging data, and seismic data; the target parameters include well depth and its corresponding formation velocity; Extracting well depth data and corresponding interval velocity data corresponding to the target parameter from the feature data, preprocessing the well depth data and the corresponding interval velocity data to obtain preprocessed well depth data and the corresponding interval velocity data, and training a pre-constructed initial LSTM model using the preprocessed well depth data and the corresponding interval velocity data to obtain an interval velocity prediction model; A multidimensional data set is constructed using the well depth data and its corresponding interval velocity data, and a formation pressure prediction model is trained by combining the calculation formula of the Eaton model; The measured well depth values of the measured points of the undrilled section located in the same address area as the drilled section are collected in real time, and the formation pressure of the undrilled section is predicted in combination with the layer velocity prediction model and the formation pressure prediction model to obtain the predicted formation pressure value of the undrilled section.
2. The method according to claim 1, characterized in that Determining a target parameter closely related to formation pressure from the characteristic data includes: Performing Spearman correlation analysis on the well logging data, the mud logging data, and the seismic data to calculate the rank correlation coefficient between each parameter in the well logging data, the mud logging data, and the seismic data and the formation pressure; The parameter corresponding to the rank correlation coefficient greater than the preset rank value is determined as the target parameter.
3. The method according to claim 1, characterized in that The preprocessing of the well depth data and the corresponding interval velocity data to obtain the preprocessed well depth data and the corresponding interval velocity data includes: performing data cleaning on the well depth data and the corresponding interval velocity data to obtain cleaned well depth data and the corresponding interval velocity data; Normalization processing is performed on the cleaned well depth data and the corresponding interval velocity data to obtain pre-processed well depth data and the corresponding interval velocity data.
4. The method according to claim 1, wherein The pre-constructed initial LSTM model is trained using the pre-processed well depth data and its corresponding interval velocity data to obtain an interval velocity prediction model, including: The preprocessed well depth data and its corresponding interval velocity data are divided into a training set and a test set; The initial LSTM model is iteratively trained using the training set until the training is stopped when a preset number of iterations is reached, thereby obtaining a corresponding layer velocity prediction model.
5. The method according to claim 4, characterized in that After the step of iteratively training the initial LSTM model using the training set until a preset number of iterations is reached and the training is stopped to obtain the corresponding layer velocity prediction model, the method further includes: Inputting the true value of the well depth in the test set into the interval velocity prediction model, and outputting the interval velocity test value corresponding to the true value of the well depth; Obtaining a true value of interval velocity corresponding to the true value of the well depth from the test set, and calculating a mean absolute error and a root mean square error of the interval velocity prediction model using the true value of the interval velocity and the interval velocity test value; The model performance of the interval velocity prediction model is evaluated using the mean absolute error and the root mean square error.
6. The method according to claim 1, characterized in that The multidimensional data set is constructed by using the well depth data and the corresponding interval velocity data, and the formation pressure prediction model is trained in combination with the calculation formula of the Eaton model, including: Calculating various parameters for training the formation pressure prediction model based on the well depth data and its corresponding interval velocity data, and constructing a multidimensional data set using the various parameters; According to the multidimensional data set, the calculation formula of the Eaton model is used as a basic model for training to obtain a formation pressure prediction model.
7. The method according to claim 1, characterized in that The real-time acquisition of measured well depth values of measured points of an undrilled section located in the same address area as the drilled section, and prediction of the formation pressure of the undrilled section in combination with the interval velocity prediction model and the formation pressure prediction model to obtain a predicted formation pressure value of the undrilled section, includes: Real-time acquisition of measured well depth values of measured points in an undrilled section located in the same address area as the drilled section, inputting the measured well depth values into the interval velocity prediction model, and outputting interval velocity prediction values corresponding to the measured well depth values; According to the measured well depth value and the corresponding predicted layer velocity value, the formation pressure of the undrilled well section is predicted using the formation pressure prediction model to obtain a predicted formation pressure value.
8. A pre-drilling formation pressure prediction system, characterized in that: The system comprises: a determination module, configured to obtain characteristic data at a measured point of formation pressure in a drilled well, and determine a target parameter closely related to the formation pressure from the characteristic data; wherein the characteristic data includes well logging data, mud logging data, and seismic data; and the target parameter includes well depth and its corresponding formation velocity; a preprocessing module, configured to extract the well depth data and the corresponding interval velocity data corresponding to the target parameter from the feature data, preprocess the well depth data and the corresponding interval velocity data to obtain the preprocessed well depth data and the corresponding interval velocity data, and train a pre-built initial LSTM model using the preprocessed well depth data and the corresponding interval velocity data to obtain an interval velocity prediction model; A training module is used to construct a multidimensional data set using the well depth data and its corresponding interval velocity data, and to train a formation pressure prediction model in combination with the calculation formula of the Eaton model; The prediction module is used to collect the measured well depth values of the measured points of the undrilled section located in the same address area as the drilled section in real time, and predict the formation pressure of the undrilled section in combination with the layer velocity prediction model and the formation pressure prediction model to obtain the predicted value of the formation pressure of the undrilled section.
9. An electronic device, characterized in that: The electronic device includes: a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device performs the steps of the method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, and when the program is run, it is used to implement the steps of the method according to any one of claims 1 to 7.